Commit
•
1093ad1
0
Parent(s):
Update files from the datasets library (from 1.0.0)
Browse filesRelease notes: https://github.com/huggingface/datasets/releases/tag/1.0.0
- .gitattributes +27 -0
- dataset_infos.json +1 -0
- dummy/iwslt2017-ar-en/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-de-en/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-ar/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-de/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-fr/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-it/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-ja/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-ko/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-nl/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-ro/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-en-zh/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-fr-en/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-it-en/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-it-nl/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-it-ro/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-ja-en/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-ko-en/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-nl-en/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-nl-it/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-nl-ro/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-ro-en/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-ro-it/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-ro-nl/1.0.0/dummy_data.zip +3 -0
- dummy/iwslt2017-zh-en/1.0.0/dummy_data.zip +3 -0
- iwslt2017.py +214 -0
.gitattributes
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
4 |
+
*.bin.* filter=lfs diff=lfs merge=lfs -text
|
5 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
11 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
12 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
13 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
14 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
15 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
16 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
17 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
18 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
19 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
20 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
21 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
22 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
23 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
24 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
25 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
26 |
+
*.zstandard filter=lfs diff=lfs merge=lfs -text
|
27 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
dataset_infos.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"iwslt2017-en-it": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "it"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-it", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 46648117, "num_examples": 231619, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 305254, "num_examples": 1566, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 200031, "num_examples": 929, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 47153402, "size_in_bytes": 376484681}, "iwslt2017-en-nl": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "nl"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-nl", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 42844125, "num_examples": 237240, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 311654, "num_examples": 1777, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 197822, "num_examples": 1003, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 43353601, "size_in_bytes": 372684880}, "iwslt2017-en-ro": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "ro"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-ro", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 44130134, "num_examples": 220538, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 316798, "num_examples": 1678, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 205036, "num_examples": 914, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 44651968, "size_in_bytes": 373983247}, "iwslt2017-it-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["it", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-it-en", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 46648117, "num_examples": 231619, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 305254, "num_examples": 1566, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 200031, "num_examples": 929, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 47153402, "size_in_bytes": 376484681}, "iwslt2017-it-nl": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["it", "nl"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-it-nl", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 43033360, "num_examples": 233415, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 309733, "num_examples": 1669, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 197782, "num_examples": 1001, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 43540875, "size_in_bytes": 372872154}, "iwslt2017-it-ro": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["it", "ro"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-it-ro", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 44485345, "num_examples": 217551, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 314982, "num_examples": 1643, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 204997, "num_examples": 914, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 45005324, "size_in_bytes": 374336603}, "iwslt2017-nl-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["nl", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-nl-en", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 42844125, "num_examples": 237240, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 311654, "num_examples": 1777, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 197822, "num_examples": 1003, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 43353601, "size_in_bytes": 372684880}, "iwslt2017-nl-it": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["nl", "it"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-nl-it", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 43033360, "num_examples": 233415, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 309733, "num_examples": 1669, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 197782, "num_examples": 1001, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 43540875, "size_in_bytes": 372872154}, "iwslt2017-nl-ro": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["nl", "ro"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-nl-ro", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 41338906, "num_examples": 206920, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 320960, "num_examples": 1680, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 202388, "num_examples": 913, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 41862254, "size_in_bytes": 371193533}, "iwslt2017-ro-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["ro", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-ro-en", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 44130134, "num_examples": 220538, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 316798, "num_examples": 1678, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 205036, "num_examples": 914, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 44651968, "size_in_bytes": 373983247}, "iwslt2017-ro-it": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["ro", "it"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-ro-it", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 44485345, "num_examples": 217551, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 314982, "num_examples": 1643, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 204997, "num_examples": 914, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 45005324, "size_in_bytes": 374336603}, "iwslt2017-ro-nl": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["ro", "nl"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-ro-nl", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 41338906, "num_examples": 206920, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 320960, "num_examples": 1680, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 202388, "num_examples": 913, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz": {"num_bytes": 329331279, "checksum": "cebb3839b7580f8e4eecf66d75424caa29b9410a9a491a732b2d343626ee0243"}}, "download_size": 329331279, "dataset_size": 41862254, "size_in_bytes": 371193533}, "iwslt2017-ar-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["ar", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-ar-en", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 56481251, "num_examples": 231713, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 2014304, "num_examples": 8583, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 241214, "num_examples": 888, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/ar/en/ar-en.tgz": {"num_bytes": 27747532, "checksum": "f1b77caa115cf5cbd00e519b425e33615080c8735fd9dd59a0c811bacd10ab70"}}, "download_size": 27747532, "dataset_size": 58736769, "size_in_bytes": 86484301}, "iwslt2017-de-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["de", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-de-en", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 42608548, "num_examples": 206112, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1608482, "num_examples": 8079, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 210983, "num_examples": 888, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/de/en/de-en.tgz": {"num_bytes": 16751373, "checksum": "ebcd53c53c8df742437e3a0fb964772e78764c3f6d4e38df0b6a2f71abc3a8b4"}}, "download_size": 16751373, "dataset_size": 44428013, "size_in_bytes": 61179386}, "iwslt2017-en-ar": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "ar"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-ar", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 56481251, "num_examples": 231713, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 2014304, "num_examples": 8583, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 241214, "num_examples": 888, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/en/ar/en-ar.tgz": {"num_bytes": 29332212, "checksum": "7c9446ebf519fb9ca85d7fb7a6681918772eafc6a33bfdf9852f87442aabd3f7"}}, "download_size": 29332212, "dataset_size": 58736769, "size_in_bytes": 88068981}, "iwslt2017-en-de": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "de"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-de", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 42608548, "num_examples": 206112, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1608482, "num_examples": 8079, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 210983, "num_examples": 888, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/en/de/en-de.tgz": {"num_bytes": 16751364, "checksum": "623caec69988b224dbe741f16ad90411b93430117a8021c4bb2bcab7ad1eb21e"}}, "download_size": 16751364, "dataset_size": 44428013, "size_in_bytes": 61179377}, "iwslt2017-en-fr": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "fr"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-fr", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 49273478, "num_examples": 232825, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1767473, "num_examples": 8597, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 207587, "num_examples": 890, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/en/fr/en-fr.tgz": {"num_bytes": 27691674, "checksum": "1f5bd8cd242de15fb58a04af265adb8e604a7e1b805a42cb0de7dfd15447c5c6"}}, "download_size": 27691674, "dataset_size": 51248538, "size_in_bytes": 78940212}, "iwslt2017-en-ja": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "ja"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-ja", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 48205171, "num_examples": 223108, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1809015, "num_examples": 8469, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 208132, "num_examples": 871, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/en/ja/en-ja.tgz": {"num_bytes": 26983540, "checksum": "27322c1170d057a92f27edd475cfd2c46f38bd1acdd52bbcffbe725e78b3d472"}}, "download_size": 26983540, "dataset_size": 50222318, "size_in_bytes": 77205858}, "iwslt2017-en-ko": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "ko"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-ko", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 51678235, "num_examples": 230240, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1869801, "num_examples": 8514, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 219303, "num_examples": 879, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/en/ko/en-ko.tgz": {"num_bytes": 19363699, "checksum": "0ac6caa1fbfbb0c79f1e9ef2f8e5493b83003d6ad85ef504a5584e3b8b006f00"}}, "download_size": 19363699, "dataset_size": 53767339, "size_in_bytes": 73131038}, "iwslt2017-en-zh": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["en", "zh"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-en-zh", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 44271196, "num_examples": 231266, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1605535, "num_examples": 8549, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 202545, "num_examples": 879, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/en/zh/en-zh.tgz": {"num_bytes": 27596371, "checksum": "d178f2ede07d7d600e244649e2092fb5fcbdda9729a3bb016502512028e2615b"}}, "download_size": 27596371, "dataset_size": 46079276, "size_in_bytes": 73675647}, "iwslt2017-fr-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["fr", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-fr-en", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 49273478, "num_examples": 232825, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1767473, "num_examples": 8597, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 207587, "num_examples": 890, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/fr/en/fr-en.tgz": {"num_bytes": 26872507, "checksum": "a40ca3b9a0064bd1fd582df4ced260dd3e7ff9f6f7bf863f139f74b7a5cabd78"}}, "download_size": 26872507, "dataset_size": 51248538, "size_in_bytes": 78121045}, "iwslt2017-ja-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["ja", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-ja-en", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 48205171, "num_examples": 223108, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1809015, "num_examples": 8469, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 208132, "num_examples": 871, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/ja/en/ja-en.tgz": {"num_bytes": 26190272, "checksum": "19b394f0256e7398bc69fb27c0fde15cd2a6c372a910ad99707ae32d54371640"}}, "download_size": 26190272, "dataset_size": 50222318, "size_in_bytes": 76412590}, "iwslt2017-ko-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["ko", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-ko-en", "version": {"version_str": "0.0.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 51678235, "num_examples": 230240, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1869801, "num_examples": 8514, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 219303, "num_examples": 879, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/ko/en/ko-en.tgz": {"num_bytes": 19363701, "checksum": "c0cba8c055a9bd55902aff5d61783ddad281df86f2ebb673d97b126b734a0987"}}, "download_size": 19363701, "dataset_size": 53767339, "size_in_bytes": 73131040}, "iwslt2017-zh-en": {"description": "The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:\n\n German, English, Italian, Dutch, Romanian.\n\nFor each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.\n", "citation": "@inproceedings{cettoloEtAl:EAMT2012,\nAddress = {Trento, Italy},\nAuthor = {Mauro Cettolo and Christian Girardi and Marcello Federico},\nBooktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},\nDate = {28-30},\nMonth = {May},\nPages = {261--268},\nTitle = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},\nYear = {2012}}\n", "homepage": "https://sites.google.com/site/iwsltevaluation2017/TED-tasks", "license": "", "features": {"translation": {"languages": ["zh", "en"], "id": null, "_type": "Translation"}}, "supervised_keys": null, "builder_name": "iwsl_t217", "config_name": "iwslt2017-zh-en", "version": "0.0.0", "splits": {"train": {"name": "train", "num_bytes": 44271196, "num_examples": 231266, "dataset_name": "iwsl_t217"}, "test": {"name": "test", "num_bytes": 1605535, "num_examples": 8549, "dataset_name": "iwsl_t217"}, "validation": {"name": "validation", "num_bytes": 202545, "num_examples": 879, "dataset_name": "iwsl_t217"}}, "download_checksums": {"https://wit3.fbk.eu/archive/2017-01-trnted/texts/zh/en/zh-en.tgz": {"num_bytes": 26848340, "checksum": "8cecd4e8e196e5cd1e0af48b30e26c36acd33002d1e0f6e448570f68da290258"}}, "download_size": 26848340, "dataset_size": 46079276, "size_in_bytes": 72927616}}
|
dummy/iwslt2017-ar-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d72e7d76f4fcc1f7c65bff24dc0c601786239d40ea7a9ff9e223284bef441b91
|
3 |
+
size 28959
|
dummy/iwslt2017-de-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d6c937c25e1c1070f69aa92a768afc3b9925cd61ae95c64d9cb7412c0f1edf90
|
3 |
+
size 26978
|
dummy/iwslt2017-en-ar/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8a1a9cdba76ff0b24aa728d024a7fa6786144744773552a91afcdfb9f3a3498d
|
3 |
+
size 28966
|
dummy/iwslt2017-en-de/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6803837cedf33387ab5dce24342f12c2c94ae44ea464010011721af74ecbe6a0
|
3 |
+
size 26985
|
dummy/iwslt2017-en-fr/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e89da95a97f1d75be5d16ee5ef248f0c7796076455636252d2257038c0b48e84
|
3 |
+
size 27005
|
dummy/iwslt2017-en-it/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:09955c331c981409b0c297218c3c3fd99089441b7a9a1915cefce364d6d63b72
|
3 |
+
size 9840
|
dummy/iwslt2017-en-ja/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0ea1a70d4b24f366b5ab3e8c43f45e6ee56331e2bd06f25b6b18844816e14fac
|
3 |
+
size 28841
|
dummy/iwslt2017-en-ko/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f8a616ca969db95f8a7fc9abdda412d31852d141a970f58337de61a2c9feaf04
|
3 |
+
size 27504
|
dummy/iwslt2017-en-nl/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:89402df8d1d689cd96103f401220a0216cb7c248ee6e6a2d49bd0702d8b7cf98
|
3 |
+
size 10254
|
dummy/iwslt2017-en-ro/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:73ac95b8ca81c2b7411a564eb9663b41ecd669f058cd752bd297709f4a21691c
|
3 |
+
size 10554
|
dummy/iwslt2017-en-zh/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8168cd4a0354a547377318f55308fe1760b04271f806bc31e0e7d31d444bdb06
|
3 |
+
size 25303
|
dummy/iwslt2017-fr-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:17a332aa25db413ada3b023be29274afcfa9933c9d39fdaf9be52775e8912bb6
|
3 |
+
size 26998
|
dummy/iwslt2017-it-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3515f9f7731adfc434898b066b88e37d759bc5183d6c0ff00542616a183f124c
|
3 |
+
size 9840
|
dummy/iwslt2017-it-nl/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e4bf2341e18f6c01bc86c9c81dae6fb7b4ff569075926a6c9db849dc22ccd828
|
3 |
+
size 10878
|
dummy/iwslt2017-it-ro/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7d625509ea821bf1064aba584217f82c678ca9a99618be5d452445f5e6d6c8cf
|
3 |
+
size 11162
|
dummy/iwslt2017-ja-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:86fe3c437924c9dadf344ae7a5128f4b9d13f14df838efd16261835c40f6b4de
|
3 |
+
size 28848
|
dummy/iwslt2017-ko-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:51875ac35365b50a42403ce96c89ae49c9f9f38c0e57d0f8b5ca42203c983a7b
|
3 |
+
size 27497
|
dummy/iwslt2017-nl-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:91f2f280ea24171448b7eead64e6e17df0d9b3eb321c3071f190363d3836d68c
|
3 |
+
size 10250
|
dummy/iwslt2017-nl-it/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ae95c6ff98712312e208508ee3e7b1ff8a67795d0cfb023a6bef0084e59429aa
|
3 |
+
size 10874
|
dummy/iwslt2017-nl-ro/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:56f3a1aa7f7dcf24e5525bdfb15642ca5a8c8c985b14430a8b2f869eb752e4bf
|
3 |
+
size 10922
|
dummy/iwslt2017-ro-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1adee956b58dbb3963688ce74a153596f66f3364302c25dd5279c8b2024d34a1
|
3 |
+
size 10556
|
dummy/iwslt2017-ro-it/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6932f9e3a4b7d671f1a04a3d3da9637ddb00840c4c5cadcbfbbf5f2cd6b11b18
|
3 |
+
size 11164
|
dummy/iwslt2017-ro-nl/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a33a2eaa429bf7878c90bce3b5857c1d0dd7bd71481312a551f1e115d6bbb980
|
3 |
+
size 10928
|
dummy/iwslt2017-zh-en/1.0.0/dummy_data.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:738c269897145efef7686164c74d870748e8619af1224f3ff45478efdfcf7626
|
3 |
+
size 25303
|
iwslt2017.py
ADDED
@@ -0,0 +1,214 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# coding=utf-8
|
2 |
+
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
3 |
+
#
|
4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
5 |
+
# you may not use this file except in compliance with the License.
|
6 |
+
# You may obtain a copy of the License at
|
7 |
+
#
|
8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
9 |
+
#
|
10 |
+
# Unless required by applicable law or agreed to in writing, software
|
11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
+
# See the License for the specific language governing permissions and
|
14 |
+
# limitations under the License.
|
15 |
+
"""IWSLT 2017 dataset """
|
16 |
+
|
17 |
+
from __future__ import absolute_import, division, print_function
|
18 |
+
|
19 |
+
import os
|
20 |
+
|
21 |
+
import datasets
|
22 |
+
|
23 |
+
|
24 |
+
_CITATION = """\
|
25 |
+
@inproceedings{cettoloEtAl:EAMT2012,
|
26 |
+
Address = {Trento, Italy},
|
27 |
+
Author = {Mauro Cettolo and Christian Girardi and Marcello Federico},
|
28 |
+
Booktitle = {Proceedings of the 16$^{th}$ Conference of the European Association for Machine Translation (EAMT)},
|
29 |
+
Date = {28-30},
|
30 |
+
Month = {May},
|
31 |
+
Pages = {261--268},
|
32 |
+
Title = {WIT$^3$: Web Inventory of Transcribed and Translated Talks},
|
33 |
+
Year = {2012}}
|
34 |
+
"""
|
35 |
+
|
36 |
+
_DESCRIPTION = """\
|
37 |
+
The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task. The languages involved are five:
|
38 |
+
|
39 |
+
German, English, Italian, Dutch, Romanian.
|
40 |
+
|
41 |
+
For each language pair, training and development sets are available through the entry of the table below: by clicking, an archive will be downloaded which contains the sets and a README file. Numbers in the table refer to millions of units (untokenized words) of the target side of all parallel training sets.
|
42 |
+
"""
|
43 |
+
|
44 |
+
MULTI_URL = "https://wit3.fbk.eu/archive/2017-01-trnmted//texts/DeEnItNlRo/DeEnItNlRo/DeEnItNlRo-DeEnItNlRo.tgz"
|
45 |
+
|
46 |
+
|
47 |
+
class IWSLT2017Config(datasets.BuilderConfig):
|
48 |
+
""" BuilderConfig for NewDataset"""
|
49 |
+
|
50 |
+
def __init__(self, pair, is_multilingual, **kwargs):
|
51 |
+
"""
|
52 |
+
|
53 |
+
Args:
|
54 |
+
pair: the language pair to consider
|
55 |
+
is_multilingual: Is this pair in the multilingual dataset (download source is different)
|
56 |
+
**kwargs: keyword arguments forwarded to super.
|
57 |
+
"""
|
58 |
+
self.pair = pair
|
59 |
+
self.is_multilingual = is_multilingual
|
60 |
+
super().__init__(**kwargs)
|
61 |
+
|
62 |
+
|
63 |
+
# XXX: Artificially removed DE from here, as it also exists within bilingual data
|
64 |
+
MULTI_LANGUAGES = ["en", "it", "nl", "ro"]
|
65 |
+
BI_LANGUAGES = ["ar", "de", "en", "fr", "ja", "ko", "zh"]
|
66 |
+
MULTI_PAIRS = [f"{source}-{target}" for source in MULTI_LANGUAGES for target in MULTI_LANGUAGES if source != target]
|
67 |
+
BI_PAIRS = [
|
68 |
+
f"{source}-{target}"
|
69 |
+
for source in BI_LANGUAGES
|
70 |
+
for target in BI_LANGUAGES
|
71 |
+
if source != target and (source == "en" or target == "en")
|
72 |
+
]
|
73 |
+
|
74 |
+
PAIRS = MULTI_PAIRS + BI_PAIRS
|
75 |
+
|
76 |
+
|
77 |
+
class IWSLT217(datasets.GeneratorBasedBuilder):
|
78 |
+
"""The IWSLT 2017 Evaluation Campaign includes a multilingual TED Talks MT task."""
|
79 |
+
|
80 |
+
VERSION = datasets.Version("1.0.0")
|
81 |
+
|
82 |
+
# This is an example of a dataset with multiple configurations.
|
83 |
+
# If you don't want/need to define several sub-sets in your dataset,
|
84 |
+
# just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.
|
85 |
+
BUILDER_CONFIG_CLASS = IWSLT2017Config
|
86 |
+
BUILDER_CONFIGS = [
|
87 |
+
IWSLT2017Config(
|
88 |
+
name="iwslt2017-" + pair,
|
89 |
+
description="A small dataset",
|
90 |
+
version=datasets.Version("1.0.0"),
|
91 |
+
pair=pair,
|
92 |
+
is_multilingual=pair in MULTI_PAIRS,
|
93 |
+
)
|
94 |
+
for pair in PAIRS
|
95 |
+
]
|
96 |
+
|
97 |
+
def _info(self):
|
98 |
+
return datasets.DatasetInfo(
|
99 |
+
# This is the description that will appear on the datasets page.
|
100 |
+
description=_DESCRIPTION,
|
101 |
+
# datasets.features.FeatureConnectors
|
102 |
+
features=datasets.Features(
|
103 |
+
{"translation": datasets.features.Translation(languages=self.config.pair.split("-"))}
|
104 |
+
),
|
105 |
+
# If there's a common (input, target) tuple from the features,
|
106 |
+
# specify them here. They'll be used if as_supervised=True in
|
107 |
+
# builder.as_dataset.
|
108 |
+
supervised_keys=None,
|
109 |
+
# Homepage of the dataset for documentation
|
110 |
+
homepage="https://sites.google.com/site/iwsltevaluation2017/TED-tasks",
|
111 |
+
citation=_CITATION,
|
112 |
+
)
|
113 |
+
|
114 |
+
def _split_generators(self, dl_manager):
|
115 |
+
"""Returns SplitGenerators."""
|
116 |
+
source, target = self.config.pair.split("-")
|
117 |
+
if self.config.is_multilingual:
|
118 |
+
dl_dir = dl_manager.download_and_extract(MULTI_URL)
|
119 |
+
data_dir = os.path.join(dl_dir, "DeEnItNlRo-DeEnItNlRo")
|
120 |
+
years = [2010]
|
121 |
+
else:
|
122 |
+
bi_url = f"https://wit3.fbk.eu/archive/2017-01-trnted/texts/{source}/{target}/{source}-{target}.tgz"
|
123 |
+
dl_dir = dl_manager.download_and_extract(bi_url)
|
124 |
+
data_dir = os.path.join(dl_dir, f"{source}-{target}")
|
125 |
+
years = [2010, 2011, 2012, 2013, 2014, 2015]
|
126 |
+
return [
|
127 |
+
datasets.SplitGenerator(
|
128 |
+
name=datasets.Split.TRAIN,
|
129 |
+
# These kwargs will be passed to _generate_examples
|
130 |
+
gen_kwargs={
|
131 |
+
"source_files": [
|
132 |
+
os.path.join(
|
133 |
+
data_dir,
|
134 |
+
"train.tags.{}.{}".format(self.config.pair, source),
|
135 |
+
)
|
136 |
+
],
|
137 |
+
"target_files": [
|
138 |
+
os.path.join(
|
139 |
+
data_dir,
|
140 |
+
"train.tags.{}.{}".format(self.config.pair, target),
|
141 |
+
)
|
142 |
+
],
|
143 |
+
"split": "train",
|
144 |
+
},
|
145 |
+
),
|
146 |
+
datasets.SplitGenerator(
|
147 |
+
name=datasets.Split.TEST,
|
148 |
+
# These kwargs will be passed to _generate_examples
|
149 |
+
gen_kwargs={
|
150 |
+
"source_files": [
|
151 |
+
os.path.join(
|
152 |
+
data_dir,
|
153 |
+
"IWSLT17.TED.tst{}.{}.{}.xml".format(year, self.config.pair, source),
|
154 |
+
)
|
155 |
+
for year in years
|
156 |
+
],
|
157 |
+
"target_files": [
|
158 |
+
os.path.join(
|
159 |
+
data_dir,
|
160 |
+
"IWSLT17.TED.tst{}.{}.{}.xml".format(year, self.config.pair, target),
|
161 |
+
)
|
162 |
+
for year in years
|
163 |
+
],
|
164 |
+
"split": "test",
|
165 |
+
},
|
166 |
+
),
|
167 |
+
datasets.SplitGenerator(
|
168 |
+
name=datasets.Split.VALIDATION,
|
169 |
+
# These kwargs will be passed to _generate_examples
|
170 |
+
gen_kwargs={
|
171 |
+
"source_files": [
|
172 |
+
os.path.join(
|
173 |
+
data_dir,
|
174 |
+
"IWSLT17.TED.dev2010.{}.{}.xml".format(self.config.pair, source),
|
175 |
+
)
|
176 |
+
],
|
177 |
+
"target_files": [
|
178 |
+
os.path.join(
|
179 |
+
data_dir,
|
180 |
+
"IWSLT17.TED.dev2010.{}.{}.xml".format(self.config.pair, target),
|
181 |
+
)
|
182 |
+
],
|
183 |
+
"split": "dev",
|
184 |
+
},
|
185 |
+
),
|
186 |
+
]
|
187 |
+
|
188 |
+
def _generate_examples(self, source_files, target_files, split):
|
189 |
+
""" Yields examples. """
|
190 |
+
id_ = 0
|
191 |
+
source, target = self.config.pair.split("-")
|
192 |
+
for source_file, target_file in zip(source_files, target_files):
|
193 |
+
with open(source_file, "r", encoding="utf-8") as sf:
|
194 |
+
with open(target_file, "r", encoding="utf-8") as tf:
|
195 |
+
for source_row, target_row in zip(sf, tf):
|
196 |
+
source_row = source_row.strip()
|
197 |
+
target_row = target_row.strip()
|
198 |
+
|
199 |
+
if source_row.startswith("<"):
|
200 |
+
if source_row.startswith("<seg"):
|
201 |
+
# Remove <seg id="1">.....</seg>
|
202 |
+
# Very simple code instead of regex or xml parsing
|
203 |
+
part1 = source_row.split(">")[1]
|
204 |
+
source_row = part1.split("<")[0]
|
205 |
+
part1 = target_row.split(">")[1]
|
206 |
+
target_row = part1.split("<")[0]
|
207 |
+
|
208 |
+
source_row = source_row.strip()
|
209 |
+
target_row = target_row.strip()
|
210 |
+
else:
|
211 |
+
continue
|
212 |
+
|
213 |
+
yield id_, {"translation": {source: source_row, target: target_row}}
|
214 |
+
id_ += 1
|